Forecasting Method of epileptic seizures by using the coercive adjusted auto-regressive model

摘要

PURPOSE: An epileptic seizure predicting method using an adaptive autoregressive model of mandatory access is provided to increase an epileptic seizure prediction rate of epileptic brainwave data having complex features such as non-linear, informal, and aperiodic features. CONSTITUTION: An autoregressive model is established to estimate a future value by using data values. The autoregressive model is established by setting an autoregressive parameter of the model as '1'. A fractal dimension is used for determining a degree of the autoregressive parameter and a box-counting method is applied for calculating the fractal dimension.